Research And Application Of Fractional Particle Swarm Optimization Algorithm

Particle Swarm Optimization(PSO)algorithm is a new swarm intelligence optimization algorithm inspired by the foraging phenomenon of biological groups such as birds and fish.It has developed rapidly in recent years and has been widely used in many fields.It has the advantages of a simpler concept,easier implementation,faster convergence speed,and fewer parameters to be adjusted in the algorithm.However,the PSO algorithm still faces problems such as low search accuracy,slow local convergence speed,easy to fall into local optimum for complex multimodal problems,and uneven distribution of the obtained solutions.Therefore,it has attracted many domestic and foreign scholars to carry out a lot of improvement research on it.In this thesis,the existing fractional PSO algorithm has a strong dependence on a single fractional operator in the iterative process: when the value of the fractional operator increases,the convergence speed of the population becomes slower,and when the value of the fractional operator decreases When the population is small,the probability of the population falling into the local optimum becomes high.An improved fractional particle swarm optimization(IFPSO)is proposed.The improved algorithm introduces a new fractional operator and adopts a linear decreasing strategy for the inertia weight factor,and constructs a new update formula of particle swarm optimization.The comparison results of multiple sets of test functions show that the new optimization algorithm enhances the diversity of particles,showing the controllability of local convergence and the high efficiency of global convergence.Improves the ability of the algorithm to capture the global optimal solution.The IFPSO algorithm is applied to the parameter optimization of support vector machine and K-means algorithm in this thesis.The improved algorithm is used to solve the constrained optimization problem,and the single-class and multi-class prediction models are constructed by encapsulation with the support vector machine,which effectively improves the prediction accuracy of the model.The cluster center of Kmeans is optimized to improve the convergence of the algorithm and the accuracy of the optimal solution.The effectiveness and stability of the IFPSO algorithm are verified by simulation experiments.